[Accessibility] KDEs position towards CORBA (was: draft Accessibility meeting minutes 9/15/04)

[Gunnar Schmi Dt]

On Wednesday 15 September 2004 21:15, John Goldthwaite wrote:
> Accessibility meeting 9/15/04
> [...]
> Bill- there some issues there that are appropriate for discussion within
> the group but I don’t think some are necessarily representative of the
> group. If we felt that all of the topics such AT-SPI and DBUS should be
> published, I think that Harald and I could get together and clean it up.
>  The other possibility is to just pull the section.
> [...]

Actually simply removing this part from the minutes does not seem right to 
me as using CORBA as a base for AT-SPI is problematic for KDE based AT 
clients.

> Bill- these are some of the issues that the group talked about a year
> ago. In their current form, the minutes are not helpful even as a record
> on the meeting.
> [...]

I have looked into the minutes in order to refresh my memory of these 
discussions:

On February 4, 2004, I had voiced some concerns that using CORBA could be a 
problem for KDE based AT clients. (Unfortunately my initial statement in 
that meeting is cut off in the middle, so that it makes the impression as 
if the bridge from the Qt Accessibility Framework would solve both 
directions. However, this bridge is only useful for the server side.)

Some discussions followed whether using DCOP or DBUS instead of CORBA is 
possible. However, there was no conclusion about that.

On February 11, 2004, I was not on the meeting. According to the minutes it 
was decided to defer the discusion about the low-level IPC protocol that 
AT-SPI is based on.

During the following two meetings some aspects about CORBA were discussed 
as AT-SPI is currently based on CORBA. Part of the results of these 
discussions were that for speed reasons it is necessary to use ORBIT2 (or 
a compatible ORB, which to my knowledge does not exist). This would mean 
that all KDE based AT clients would need to integrate two event loops into 
each other (both the glib and Qt have their own event loop), which sounds 
to be non-trivial to me.


I do not want to sabotage the AT-SPI subgroup, but it seemed to be 
necessary to me to point out that the problem of KDE based AT clients is 
not yet solved. While it is an important step to create a standard like 
AT-SPI we should not make something a standard that creates unnecessary 
difficulties for writing AT-clients based on the technologies of either of 
the two big desktops.

Gunnar Schmi Dt
-- 
Co-maintainer of the KDE Accessibility Project
Maintainer of the kdeaccessibility package
http://accessibility.kde.org/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.linux-foundation.org/pipermail/accessibility/attachments/20040916/8d443e1d/attachment.pgp